Found the paper again
AES no.1324 “Low Distortion Loudspeaker System”. 37th convention 1969...
Earl Geddes should read the introduction.
Otherwise, thanks for posting. How much have we progressed since 1969!
This is 100 dB at 1 meter anechoic. Smallish speaker <6 kg.
A lot of what makes a speaker sound better than another is simply less resonance. Electrostats and other bipoles don't have as much inherent resonance issues. They have other issues though.
Another major issue is the phase distortion that results when you have sound coming from two points in space greater than 1/2 wavelength apart. This creates off axis anomolies. Also driver beaming contributes to this.
The solution here is simple, use equalizer APO and do a "cliff slope" crossover. This will get rid of basically all lobing and phase issues that afflict multi driver speakers. It also reduces the distortion and beaming that results from woofers and tweeters playing out of their comfort zone.
Finally a steep slope gets rid of the "cone cry" break up of hard cones. Done, problem solved. You still need a well built cabinet that doesn't resonate and reduces diffraction and reflections from the tweeter.
Oh and you can equalize out the linear distortion, which is particularly annoying in the 1-8 khz region.
Last edited:
Distortion does become more audible when you minimize resonances and decay. I was with someone whom had been involved in many aspects of music reproduction listening to my setup. Used some loudness without calibrating it and he did point out some issues at the lower frequency during a piano session. We both had the same view that the impact of a live piano was not there, the harmonics and the clarity was pretty good.
You really want to be less than a quarter wavelenght apart if you want to be able to mostly ignore the disparitiate sources.
dave
A lot of distortion is resonance. Non-linear is inherent to the driver itself, and can'r be gotten rid of. But it's not particularly noticeable either, and goes up a lot depending on volume.
Last edited:
Yes I should have said 1/4 wavelength, though I don't know that it becomes instantly noticeable at that point.
I think if you have been able to lower CSD faster, you might discover something different. That was what I did. Then suddenly I discovered more work was necessary.
Consider "psychoacoustics". A musical instrument's spectrum typically has low amplitude of the fundamental so if the harmonics' amplitudes are generally higher- up to 40 dB- than the fundamental why does the instrument NOT sound "distorted"? How is one musical instrument distinguished from another? The objective- "scientific" - answers to these questions are long known. These must be the basis of discussion of distortion.
I don't have that paper, and being so old it's not worth buying. Could you tell me what the introduction says?
Aside from the El Greco fallacy that usually shows up when a thread turns to psychology....
the answer is: the trumpet sounds like a trumpet.*
Depending on the assortment of harmonics and other sounds, an instrument may seem kind of shrill (like reed stops on an organ) or kind of mellow (like flue stops on an organ).
But I can't say as I really understand why or how the brain creates a virtual or implied fundamental tone. Basilar membrane resonance? Perceptual inference? Learning?
B.
*there is another body of study having to do with identifying sounds; for example, your Dad has the same identity even if he has a terrible sore throat and even when replayed in a tapped "horn" sub.
Last edited:
or to put it another way, you try to reproduce the fundamental but it sounds like a trumpet instead, you'd call that distorted.
That after much research they found that harmonic distortion in loudspeakers was highly audible. Therefore, reducing it was a lofty goal they had set out to do.
Exactly my point of view as explained before.
Exactly my point of view as explained before.
I have focus working on a driver. Initially the idea was some some 6” driver, but some more assessment of SPL, it seem I can reduce it to about 5”, a size which is about what Ted Jordan had mentioned to be pretty much maximum recommended for a full range driver. Looking at the analysis data, it seems what Ted had mentioned seem quite correct. In the mean time, I am also looking at what can be done to reduce inductance of the motor to reduce distortion and increase power response. From the enclosure side, I am also looking at ways to reduce enclosure internal resonances, minimize the impedance peaks. There has to be a good solution in each of these areas.
For enclosure, I did have some good results using some soft washing balls dumped into the enclosure replacing the traditional stuffing, this brought out a more realistic ambient from the recording, making DAC differences more obvious. Still lots of work to be done.
For enclosure, I did have some good results using some soft washing balls dumped into the enclosure replacing the traditional stuffing, this brought out a more realistic ambient from the recording, making DAC differences more obvious. Still lots of work to be done.
Except that the most recent studies have not shown this. Apparently the problem has been solved.
Soongsc, shortening ring over the pole piece is the way to go to reduce inductance, but at the cost of some efficiency (air gap has to be wider to fit the ring).
There are a few other design issues that need to be explored. Looking at past data, I see some noise in the lower level comparing against one without a shorting ring, but I need to go through all the test equipment and reinvestigate.
How to read a speaker's Cumulative Spectral Decay plot | The Audio Annex
According to this article there is a strong correlation between dynamics and CSD. Basically you want quick energy transfer and low energy storage and resonance.
Apparently using something like an equalizer isn't a cure all to a resonance since it still rings at that frequency.
Linkwitz has an article about CSD but its not very readable for a layperson.
Issues in speaker design - 2
According to this article there is a strong correlation between dynamics and CSD. Basically you want quick energy transfer and low energy storage and resonance.
Apparently using something like an equalizer isn't a cure all to a resonance since it still rings at that frequency.
Linkwitz has an article about CSD but its not very readable for a layperson.
Issues in speaker design - 2